I'm going to throw in my two cents as a medical student. There's no way that I could give you a lucid digestion of whether or not this paper has any significance after studying medicine for thousands to tens of thousands of hours. These topics are highly nuanced and incredibly specific to a certain domain of medicine/virology.

Your point on disinformation is also salient. I've noticed that many followers on here are intelligent, but being able to synthesize the Wikipedia article gives one the basics on a topic. There are medical theories and philosophy that contradict normal thinking, but are the fundamentals of practice. I've engaged multiple users on another account about medical principles but was shocked at how resilient people are in their claims when they seriously understand math/statistics but don't understand the medical application of those statistics; the example I can think of is I know what ARDS is, the treatment guidelines for it, how intense the nursing care is, but how many people here have ever watched someone be intubated, let alone do it themselves? There are a lot of intangibles that aren't directly written in a Wikipedia article or cannot be fully realized without actually experiencing them.

I remember learning about the 10,000 hour rule, in that spending 10,000 hours doing something one becomes a master of that domain. After my experience I can say that is not the case with medicine, which has shown me serious humility.

> but was shocked at how resilient people are in their claims when they seriously understand math/statistics but don't understand the medical application of those statistics;

> but how many people here have ever watched someone be intubated, let alone do it themselves?

I am not sure I understand the connection you are trying to make here. Are you saying that just because someone understands statistics they cannot understand medical application of statistics b/c they have no experience physically working in the field? Just looking for a little more clarification on the point you are making.

I am genuinely interested in learning about multi-disciplinary disconnects in real life, because these seem to be the some of the biggest problems we face in progress today.

I want to start by saying, I appreciate your curiosity :)

I think my point above was talking about COVID. The specific example was where I was describing how at scale COVID was going to be a massive problem and have an increasing case fatality ratio as the virus spread. The reason for that is because ~19% of people, I'm probably out of date here and frankly don't feel like looking at Uptodate, go into ARDS. ARDS usually requires high flow oxygen therapy or mechanical ventilation. High flow oxygen essentially means you're hospitalized on a standard medicine floor costing ~$5k per night whereas mechanical ventilation is $10k with crucial nurses managing. A tertiary care hospital in your local probably has between 50-200 ICU beds. Rural hospitals will have 10-30.

Now the the disconnect here is with the statistics. Early estimates were putting COVID at 2-4% CFR because when you have practically unlimited resources you can save the overwhelming majority of patients. But as incident rate increases those supplies become saturated. ARDS is an intense symptomology to treat, this isn't just a cold and it is a tonne of resources. Also, the recovery times from pneumonia that leads to ARDS is usually measured in the months from initial hospitalization. Then on top of all of this COVID has a reproducibility number of 2.5-2.7, but now we're thinking it's actually around 3. Influenza is ~1.3 for reference.

So the key points of why I was afraid: 1) ARDS is life threatening 2) ARDS requires intense nursing/intensivist (ICU doc) care 3) ARDS takes months to recover from 4) COVID has a crazy reproducibility number

Here on Hacker News there are many mathematicians and statisticians who have done centuries of work, so I was told that making an assumption this early was a fools errand because the numbers were not painting the picture that I was describing: a very bad pandemic. I think I was told that because the consequences of a viral pneumonia are not common knowledge and there is a lack understanding of what entails a viral pneumonia. But the main crux of my argument was that the CFR was going to rise from the reported 2-4% all the way up to 5-10% based on the percentage of complications that were occurring. In Wuhan the CFR is 5.9% and there's still 2-3k people on ventilators months later (on top of the fact that in the American medical community there's accusations of a manipulation of those numbers....). The CFR in Italy and France at ~10% at the moment, if not higher.

I hope this explains the entire crux of what my point was earlier. I didn't want to get into too many specifics so as not to offend anyone.

It’s like if someone came to you and asked you to deploy some software so that if someone calls 911 from a cellphone their location shows up on the dispatcher’s terminal. In 2006.

“It’s just software” they say, “I’ve already written it, how hard can it be to deploy it? It’s only a few thousand lines of code!”

There has never been any "10,000 hour rule". In his book Outliers, Malcolm Gladwell misinterpreted and oversimplified the research into learning and training.

Honestly, I never really thought that the 10,000 hour thing was that lucid to begin with. Certain things are mastered in hundreds of hours while others are "never truly mastered". When I say I learned of it, it was reading the New Yorker article and then hearing it from my favorite kind of philosopher: stoned ones.

Interesting. Here's the response direct from the authors of the study that Gladwell referenced:

https://pdfs.semanticscholar.org/b0e7/19584fc4cad4487bf614c0...

That was a great read, thank you for sharing!

I agree that HN is not capable of digesting this study reasonably not meaningfully. I question using Twitter for anything meaningful though.

Twitter is often extremely useful for scientific fields, and in many ways is perfect for it. You can post a figure and legend, tag people that may find it interesting, and start threaded discussions. It's all in the public, with no invitations or closed doors, so it is open to discovery by that dozen or two dozen people in the world that may be interested.

Science is the ultimate long-tail of super-specialized interests with extremely small audiences, and Twitter is perfect for connecting those people with extremely esoteric interests.

There is a lot of crap, but science Twitter is also the only reason I'm on it. Sadly, my own biases against Twitter Kept me off it for 8+ years, but after I started finding the right people to follow, a whole new world opened up, and it is now a vital tool for me to navigate the literature and learn quickly which new tools are best.

>There is a lot of crap, but science Twitter is also the only reason I'm on it. Sadly, my own biases against Twitter Kept me off it for 8+ years, but after I started finding the right people to follow, a whole new world opened up, and it is now a vital tool for me to navigate the literature and learn quickly which new tools are best.

That's really interesting, I've been avoiding Twitter for the same reason, but I can see how it'd help keep up with fast-moving niche fields. Maybe I'll give it a try, thanks!

How does it compare with following the same account on Facebook?

I don't know of anybody that uses Facebook for science discussion, but I did know many that were using Twitter for science, which is what finally drew me in.

Facebook seems like a much worse fit than even Twitter for discussion, but I would potentially check it out if somebody I trusted had found it useful.

There are plenty of blue checkmark specialists in these fields on Twitter who are ready to go to critique and contextualize their field. Of course the issue on Twitter is A) finding them B) filtering the low effort and bad faith trolling in the replies.

Other than twitter, I'm not sure where else you can find that audience that is accessible to a layman.

Slightly off topic but I once had a question at work on a very niche corporate finance topic. So I logged onto Twitter and tagged Aswath Damodaran in the question and had the world's foremost authority in Corporate Finance respond in like 30 minutes with a detailed answer. Twitter can be amazing for finding specialized expertise if you know where to look.

https://en.wikipedia.org/wiki/Aswath_Damodaran

That is indeed the magic of Twitter. I’ve mostly avoided it since the last presidential election but prior to that my ability to connect with people doing interesting things all over the world was incredible.

Ok that’s very powerful. But yes, after 2016 I assume such places are a plaque themselves.

I don't know, I find pseudo-anonymized, but (mostly) civilized and on-point debate that touches varying spheres of sciences as well as degrees of accuracy on them, quite valuable, too. It could probably be more "valuable", if all posts were anonymized, but then it might entail all the dangers that 4chan reminiscences about.

Sometimes, reputation blocks science too, e.g. applied in reverse, the reputation-seeking prevents many false-hypothesis confirmations, because they're less likely to propel a career (or damage it).

EDIT: ...come to think of it, the scientific method is in itself reputation-agnostic.

EDIT2: ...imagine a world, where all science was public, open, but anonymized. No patents, no proprietary research. It's a dystopia for some, and possibly a utopia for others.

> The only people who I would trust are named specialists in this field, that are willing to stake bits of their reputation on what they say.

That is basically the antithesis of the enlightenment and the whole idea of the scientific method. You should not trust experts because they say so, you should trust because they are able to present evidence which is persuasive and reproducible. Even experts are wrong, and the history of the N95 mask illustrates this in the case of a relatively unknown Wu Lien-teh being correct over the objections of the prominent French Mesny. In less contemporaneous things, it took a long time in North America for gluten to be recognized as the cause of Celiac's disease because too many people there trusted the expertise of Sydney Haas who asserted that eating bananas would cure the disease.

I trust evidence and citations to evidence. Experts are generally willing and capable of providing those easily and so experts generally quickly earn my trust. Assertions based instead only on claimed or even actual expertise do not earn my trust and often arouse my suspicion.

BioBot Analytics (listed authors of the paper) is a YC company

That is very useful information, and explains why this preprint got so much traction here, I guess. I would have rather seen a post that's aimed at a more general audience, perhaps from the company blog or similar. It would have been able to provide the proper context to this audience, which is not an audience that is monitoring MedRXiv or has the skills to understand what this preprint means.

> I don't think that HN is a good venue for discussion of specialized preprints that have highly surprising results, in topics that are moving extremely quickly.

Then close the site, it is useless.

But I thing the premise is 100% wrong.

People are absolutely able to evaluate reputation and evaluate comments amongst the internet and collate them here.

Currently this as a working model is not happening for C19 though and needs to be fixed.

Or you are right HN is just a Reddit clone and pointless.